Modular Laser Launch Architecture: Analysis and Beam ... · A 100 MW launch system might have...

Kare Technical Consulting 3/23/04 1

Modular Laser Launch Architecture:Analysis and Beam Module Design

NIAC Phase I Fellows Meeting

24 March 2004

Jordin T. KareKare Technical Consulting908 15th Ave. EastSeattle, WA [email protected]

Revised 3/23/04


The Laser Launch Concept

∂

Laser and Beam Projector• Big• Heavy• Expensive• STATIONARY

Vehicle• Small• Simple• Cheap• Inert

30,000 launches per year x 100 kg = 3000 Metric tons per year!!

30,000 launches per year x 100 kg = 3000 Metric tons per year!!

Launch many small payloads on demand -- up to 10 per hour

Leave The Hard PartsOn The Ground!

Leave The Hard PartsOn The Ground!

Rule of Thumb:1 kg of payloadper MWof laser


Why Laser Launch?

• Massive launch capacity– A 100-kg launcher can put 3000 tons per year in LEO

• Very low marginal cost to orbit– Electricity, vehicle, and propellant easily <$100/lb

• Potentially low total cost to orbit– If the system is cheap enough to buy and run, and…– If there are enough payloads to launch

• Maximum safety -- no stored energy on vehicles– Enables all-azimuth launch from any site

• High reliability, easy to maintain– The hard parts stay on the ground– Vehicles are simple, mass-produced, and testable

• Ultimate launch-on-demand -- FedEx to space


Pulsed Laser Propulsion Works...


… But Has The Same Problems As Everything Else

• Development cost– Even at $10/watt, $1 Billion for 100 MW

• Technical risk– You don’t know if it will work at all without spending $$$

• In this case, for a multi-megawatt test laser

• Programmatic risk– You don’t know what it will actually cost until you’ve built it

• Big lasers have had cost/schedule/performance problems for 40years!

– Reality is always different from theory; operational systemsare always different from prototypes


The Heat Exchanger Thruster

Primary (H2) Propellant Tank

Pump(optional)

LightweightHeat Exchanger

Nozzle(s)

Laser Beam

Dense propellantinjection trades lower Ispfor higher thrust,matches exhaust velocityto vehicle velocity

Secondary(Dense)

Propellant

• Exhaust Temperature ~1000 C• Specific Impulse ~600 seconds


Heat Exchanger Thruster Advantages

• Works with any laser wavelength and pulse format• Nearly 100% efficient

– high absorption, negligible reradiation

• Simple to design– Steady flow– Simple propellant properties (especially for H2)

• Simple to build– Electroplating technique demonstrated at LLNL– Modular design scales easily to any area

• Simple to test– Works with any radiant source; doesn’t even need a laser


Current 100 MW Vehicle Concept

Stage 2 tankTotal H2 tank volume ~25 m3 Payload

Dense propellant tanks Avionics

Aeroshell

Pressurant tank

∂

~6 meters

Heat exchanger(25 m2)

∂Drop tank

∂Drop tank


What Do We Do About The #%&@ Laser?

• Lasers cost too much

– Absolute cheapest high power laser is $20-50/watt• CO2 electric discharge, with very poor beam quality

• Should scale to 100 MW, but not easily or cheaply

– Stay-on-all-day lasers above ~10 kW don’t exist

• AVLIS copper vapor lasers were 10 kW total, at a cost of

>>$1000/watt

• No one will pay to develop a large laser

– Too many bad memories: CO2, HF/DF, Excimer, FEL…

– “There are liars, damn liars, and laser builders”


Laser Diode Arrays

• >50% efficient DC to Light at ~800 nm• 10,000 hour lifetime (60,000 launches!) CW operation• Run on DC current; water cooled• Commercially available from multiple vendors

A 1200 Watt CW “stack”from Nuvonyx, Inc. --a catalog item!

A 1200 Watt CW “stack”from Nuvonyx, Inc. --a catalog item!

1 cm

• $4 - $10/watt NOW

• $2/watt in a few years in100 MW quantities

• BUT -- not coherent; nothigh enough radiance tobeam 500 km


The Beam Module Concept

• DON’T build one big laser and beam director

• Build MANY small “Beam Modules”– Completely independent laser and beam director

– Minimal common services, ideally only power and water

“This division of the laser source among many apertures was initiallyregarded only as a necessary evil, required by the low radiance ofnoncoherent [laser diode] arrays. However, we have recently realized thatthe fact that the laser and optical aperture can be subdivided into smallindependent “beam modules” is a fundamental advantage of laserpropulsion over other advanced propulsion systems, and may well be thekey to making laser launch the best option for a future launch architecture”-- J. Kare 7/03

“This division of the laser source among many apertures was initiallyregarded only as a necessary evil, required by the low radiance ofnoncoherent [laser diode] arrays. However, we have recently realized thatthe fact that the laser and optical aperture can be subdivided into smallindependent “beam modules” is a fundamental advantage of laserpropulsion over other advanced propulsion systems, and may well be thekey to making laser launch the best option for a future launch architecture”-- J. Kare 7/03


Advantages of Beam Modules

• Scalability– System grows smoothly by adding beam modules

• Reliability and maintainability– Failed modules have no effect on launch (even in progress)

– Beam modules can be replaced as units

• Cost– Everything in the system is mass-produced

– Plausible cost goal: comparable to a modern automobile(excluding laser)

• Development– All the technical risk is in the first few units: $M, not $B

– No failure costs very much -- you can’t “crash the prototype”


Conceptual Beam Module (as of last year)

Optics module• 6 kW* diode array • Stacking optics• Tip-tilt mirror• Tracking sensor

Support module• Diode power supply (16 kW* DC)• Diode temperature controller• Cooling water pump/regulator• Tracking sensor controller• Mount controller and drivers• Tip/tilt controller and drivers

Telescope• 3-m replica primary• Secondary• 2-axis alt-az mount(possibly alt-alt to avoidzenith singularity)

A 100 MW launchsystem might have

~20,000 of these* --But you can build ONE

to start with

A 100 MW launchsystem might have

~20,000 of these* --But you can build ONE

to start with

*Based on 2x1013 radiance and 500 km range;better radiance or shorter range would increaseunit power and decrease number required


At Least Three Solutions

• Fiber Lasers

• Spectral Beam Combining

• Diode Pumped Alkali Laser (DPAL)

All made breakthroughs within the last year!


Fiber Lasers

• Converts non-coherent diode array light to single-mode laseroutput with up to 90% efficiency; 75% is routine

• Demonstrated at 1 kW; 10 kW projected within 1-2 years

• Simple and mass-produceable; already in commercial production

See www.spiphotonics.com


Spectral Beam Combining (SBC)

lmax

lmin

DiffractionGrating

Diode bars

2-D Microlenses

Field lens

• Diodes operate independently in external cavity– Antireflection coating on laser diode output facets– Each diode automatically operates at the “correct” wavelength

• Demonstrated* with ~700 diodes in 7 bars (26 watt output)– >1000-fold stacking should be feasible

• SBC efficiency ~50% (power out compared to raw diode bars)

Output Coupler

•AcuLight, Inc., Bothell, WA, 2004


Diode Pumped Alkali Laser (DPAL)

• New concept (2003) developed by W. Krupke et al. (ex-LLNL)• Rb (785 nm) or Cs (895 nm) vapor in He buffer gas

– Absorption line pressure-broadened to match diode linewidth– High efficiency requires tight control of diode wavelength, spectrum

• Demonstrated at ~30 W level– Performance predicted accurately with no free parameters

4 kW 795 nm Rb laser concept• Diode array pump 6.08 kW @ 780 nm• 66% light-to-light efficiency


Laser Subsystem: Alternatives

50 kW10 kW50 kWModule output power

~300 liter/min*

162 kW

222 kW

27%

50%

18.5 kW

54%

1.1 x 1016

1.2

6

10 kW

795 nm

Rubidium vapor cell

DPAL

206# of lasers/module

600 W10 kWUnit laser power

SBC DiodesBaseline Fiber Laser

~50 liter/min*

28 kW

40 kW

30%

50%

1000 W

60% SBC eff.

2.3 x 1014

2

805 - 810 nm

8 x 125 W diode bars

~100 liter/min Water flow

90 kWCooling Requirement

150 kWDC Power (module)

40%DC efficiency (Pout / PDC)

50%Diode Efficiency

12.5 kWPump Power per laser

80%Laser Efficiency (Pout / Pdiode)

3.8 x 1015 W/m2-srRadiance (P / l2 (M2)2)

1.5Beam quality (M2)

1.08 mmWavelength

Yb-doped double-corefiber

Lasing medium


0

1

2

3

4

5

6

7

8

9

10

0.1 1.0 10.0 100.0 1000.0

Total power (MW)

$/W

att

$10/W

$6/W

Diamond Optical est.$4/W @ 1000 bars,

~20% drop per 10X qty

95% learning curve

90% learning curve

Likely price range@ 200 MW quantity ~$2 - 3.50/W

How Much Will They Cost? Diode Arrays

$/W

CurrentPrice

Range


Laser Diode Cost Trends

– Substantial reductions in bar cost• $100/bar, 100 W bars in 1-2 years (F. Way, Diamond Optical)

• $10/bar (50 W bars?) “to stay competitive” (a major manufacturer)

– Substantial reductions in packaging cost• LLNL, Oriel, others developing low-cost packaging

– Oriel “TO-220” package aimed at 50% of late-’03 price; available in mid 2004

• Diamond Optical willing to quote ~50 cents/watt

– Unpredictable gains in performance• Bars have been stuck at 60 W for several years (100 W Real Soon Now)

• Major improvement may require radical approach (e.g., VCSEL arrays)

– But...Current market does not support large investments in improvedprocessing/packaging

• ~$100M/year for all high power arrays, vs. $4B/year for discrete diodes atpeak of telecomm market

• ~$1B market for launch system would drive industry


How Much Will They Cost? Complete Lasers

• Current commercial fiber lasers are $500/watt* -- Too Much.BUT– Relatively low unit power: 100 W

– Based on discrete packaged diodes @ $100 - 150/watt, not bars

– Semi-custom production: typically ~10 units

• Best prediction for high-volume, multi-kW lasers: $7 - 10/W– 3 - 4X diode cost

– Best estimate of component cost in quantity

– Consistent with projected cost of fiber ($2K / 1 kW)

– Consistent with ~85% learning curve for complex assemblies

• Other laser options in same range


Telescope Requirements

• Total Area

– Set by laser radiance (W/m2-sr) and vehicle flux requirement• e.g., fiber laser @ 3.8 x 1015 W/m2 sr requires ~330 m2 of optics

– Secondary limits from mirror heating, thermal blooming• Still need to double check blooming

• Mirror size and quality– Diffraction: Dmirror > f l R / d

• R / d ~ 105 (500 km / 5 m) f ~ 2 (2.44 for classical limit)• ~ 16 cm @ 0.8 mm, 22 cm @ 1.08 mm

– Wavefront error: Slope < 0.5 x 10-6

• ~5 waves per meter (vs. 1/10 wave for astronomical telescope)

• Pointing and tracking– Pointing range (nominal) +/- 80° along track, +/- 45° crosstrack

– Closed loop tracking to <<10 mrad; open loop pointing to ~1 mrad


Optics Options

Cassegrain• Simple optics • Compact mount• Obscuration losses (few %)

Off-axis Cassegrain• Low loss• Asymmetric optics• Bulky

Diffractive primary• Potentially low cost primary• Potentially light weight• No obscuration• No current technology• Chromatic aberration

Stationary telescopewith tracking flat• All hardware is stationary• Minimum moving mass• Cheap primary• Additional large optic (flat)• Field rotation

Multiple small telescopeson common mount• Lower optics cost/mass• Compact• More tracking hardware• Alignment problems


Diffractionlimit ~0.2m

Productionoptics ~50 cm

Max. blank diam.0.95 - 1.4 m

Largest monolithsmade to date 8m

Optimum Primary Diameter

• Minimum size set by tracking hardware cost or diffraction

• Maximum set by rapid increase in mirror cost with size

• Small jumps at limits of various “standard” supplier capabilities

Tracking hardwareConst * N ~ D-2

0.1 m 1 m 10 m

Mirror cost~ N* D2.5

Primary Diameter

SystemCost


TANSTAACT(TAG)

• There Ain’t No Such Thing As A Cheap Telescope(That’s Any Good)– Many advanced technologies not necessary (or cheap)

• Non-glass substrates (SiC, Graphite Epoxy)• Advanced polishing techniques (magnetorheological polishing)• Active/Adaptive primary (PAMELA)

– Several possibilities still to look at• Replica optics• Electrodeposited metal optics• Lightweight mount (up to half the cost of a telescope)

• Shifts optimum toward fewer, smaller telescopes thanoriginal concept– “Only” 1000 - 2000 x 1 m, vs. 10,000 x 2 - 3 m– Allowed by improved (higher radiance) laser options


Telescope Baseline

~$100K each @ 1000 units ($22K for primary)$2 - 5M investment to be able to make 1/day

~$100K each @ 1000 units ($22K for primary)$2 - 5M investment to be able to make 1/day

• ~1 meter f/2.5 Cassegrain

– Afocal (technically a Mersenne)

– Borosilicate (Pyrex) primary• “Few-wave” accuracy

– Multilayer coated for lowabsorption

• Small (10 cm) secondary

– Minimize obscuration

– Limited field of view is OK

• Alt-Alt mount

– Tracks smoothly through zenith


Photonic Crystal Fibers

J. Limpert, et al. "Thermo-opticalproperties of air-clad photonic crystalfiber lasers in high power operation,"Opt. Express 11, 2982-2990 (2003)

• Transport kW power with low loss (<<1 dB/meter)–“Holey” fiber region guides single mode; forbids higher modes– Most power is transported in void space; avoids nonlinear effects

• Enabling for low cost beam projectors– Eliminates multiple mirrors in beam path

• Lossy, difficult to align, difficult to clean– Isolates laser from telescope motion, dust, etc....

www.blazephotonics.com

Left: “Air-clad” double core fiber

Right:Visible-wavelength high powersingle-mode guiding


Tracking Subsystem

• Requirement: point beam to ~5 mradians– Track vehicle (control mount) 500 mr @ <10 Hz– Compensate atmosphere <100 mr @ ~100 Hz

“Fast” CMOS cameradrives tip-tilt mirror;100 mr FOV, 500 Hz

“Slow” CCD cameracontrols mount;1 mr FOV, 60 Hz readout

controller

3-axis(tip/tilt/focus)actuatedmirror

From laser


Beacon Options

• Reflected laser light– Too much local scattered light

• Thermal radiation from hot heat exchanger– Start/restart problem

– No pointahead for atmospheric correction

• Ground-based laser with retroreflector– Possible, but requires high power (kW), good tracking

• Beacon on vehicle– Narrow-angle with pointing mechanism

– Wide-angle -- simplest possible system• A laser diode and a pingpong ball


Baseline Beam Module

Optics module• Fiber output optics • Tip-tilt mirror• Tracking sensor

Support module• Diode power supply (250 kW DC)• Cooling water pump/regulator• Tracking sensor controller• Mount controller and drivers• Tip/tilt controller and drivers

1-m Cassegrain Telescope• f/2.5 primary• Alt-Alt Mount

• Allows for 20% losses in optics, 10% loss intransmission, and 10% of modules offline formaintenance/repair

Laser assembly:6 x 10 kW fiber lasers


• Laser 1020– 120 MW Fiber lasers @ $8/watt 960 – 300 MW DC supply @ $.20/watt 60

• Optics 144– 2000 Primary mirrors @ $22 K 44– Other optics, pointing, tracking @ $25 K 50– Mount and pad @ $25 K 50

• Facilities 161 - 350– H2 plant (1000 - 30,000 launches/year) 2 - 60– Power buffer 15– Power line 11-48– Launch stand 30– Physical plant 100 - 195

• TOTAL 1325 - 1514

My long-standing Rule Of Thumb estimate: ~$2 Billion

100 MW System Capital Cost


What Happens To Space?Architectural Implications 1

• Cheap small payloads (common to all laser launch)1. Microsats/Nanosats: any mission that can be done in 100

kg pieces will be cheapest that way– But that will only account for a few % of launch capacity

2. Modular satellites/Constellations: Divide up functionalityinto 100 kg co-orbiting blocks (cf. NIAC work on constellations)

3. On-orbit fueling/refueling/resupply– Stimulates development of autonomous microspacecraft

rendezvous and docking, “tug” spacecraft

– Opens up high-mass-ratio mission space: Moon/Mars withstorable propellants

4. On-orbit assembly: large structures constructed on orbit

5. True space industry?


What Happens To Space Industry?Architectural Implications 2

• Routine, on-demand launch; very high reliability– Shift in spacecraft reliability criteria; “ground spares” OK

• A change in space industry– Large aerospace company resources are not required

– To build vehicles

– To build beam modules

– To build payloads

– “Learning curve” for participation is much less costly

– A change in space politics– More countries can have their own launchers, or “rent

time” on larger launchers and provide vehicles or payloads


What Happens To Human Space?Architectural Implications 3

• Immediate shift in logistics for human LEO missions– Missing/broken widgets replaceable by Next Day Space

• Scaling and reliability enable growing human presence– Laser launch is uniquely testable to ~ 10-8 failure probability

• e.g., 104 launches AND 104 abort/recoveries before flying aperson

– Initial human launch capability at TBD payload/laser power• Mercury capsule was ~1500 kg; surely we could do better?

• Potential driver for launch system growth to ~1 GW

– Growth to 2 or more person vehicle opens up passengerlaunch -- to thousands or 10’s of thousands per year


In-Space Power and PropulsionArchitectural Implications 4

• Providing electric power shifts module design goals, but“power” modules can also be used for launch– PV-compatible wavelength preferred (nominally 700 - 900 nm)*

– Higher beam quality (adaptive optics) may be desired

– However, dedicated pulsed lasers may be preferred for high-Isppulsed propulsion

• Low cost modules open design space for space power– For GEO power, each satellite can have a dedicated source

– For LEO/MEO power, modules can be distributed to many sitesworldwide

• Launcher site can provide 100-MW power levels anywhereout to GEO– Relay architectures to be explored


Beam Module Satellite Solar Power System

• Small (10 cm) optics in GEO generate practical (<1 km)spot size on Earth– Ideal application for diffractive optics?

• Optics-sized solar panel produces a convenient amountof power: ~3 W for 100 cm2

• SO... build self-contained ~10x10x10 cm beam modulesand simply stack them up to make a powersat– No high power cables

– No phase locking;• No minimum satellite size to deliver power

• Power can be shared among any number of receivers

– Modules simply clip to a frame


Beam Module SSPS

Flat filmreflectorrotates 1/daytilts 1/yearfor sun tracking

Stationary flat film reflector

Module array on simple grid

GroundPV array

Beacon/commandtransmitter

Offset fromground array

Possibility:Aerostat w.microwaveor fiber bundledownlink?

Note added 3/30/04:We have been referred to a similar SSPSconfiguration for microwaves, with a phasedarray transmitter, proposed by Nobuyuki Kaya,e.g., in Space Energy and Transportation 1 (3)1996, pp. 205-213. Modular laser configurationshave also been considered; we are still seekingdetails to compare against the concept proposedhere.


Solar Power Satellite Beam Module

PV Cell10 x 10 cm3..5 W out

Laser Driver3 W in

Electronicspower 1/2 W

Laser diode1.5 W out

Guidesensor

ControllerActuator driver

3-axis MEMSactuator

3 cmfoldmirror

Main mirror or diffractive lens ~10 cm dia


Conclusions 1: Technology Is Ready

• Lasers crossed the threshold within the last year– Performance is sufficient, and nearly certain to improve

– Costs are still high, but not inherently so• Costs will drop with volume and time

• Current optics technology is dull, but adequate– Modern glass optics are cheap enough with high-radiance lasers

• Optimum primary size is ~1 meter or less

– Innovative but unproven technologies are waiting in the wings

• No show-stoppers elsewhere in the system– Mounts, pointing and tracking, etc. are straightforward

• On-vehicle “omni” beacon looks best for pointing/tracking and makesadding adaptive optics straightforward if required

– Power storage is ripe for innovative tech (advanced batteries,flywheels) but not a system driver


Conclusions 2:Architecture Implications Are Profound

• Laser launch in general shifts paradigms– Small unit payloads, routine prompt access => on orbit industry

• Modular launcher technology changes industry– Small companies can play -- modules can come from many sources– Small countries can play -- buy their space launch “by the yard”

• Crewed flight is a new game– Continuous scaling from support (100 kg payloads) to solo

launches (~1000 kg) to taxi (tour bus?) service– Inherently high reliability, inherently testable -- tourist friendly!

• Significant effects on in-space power and propulsion– Requirements are different, but overlapping– Low-enough unit costs open new options, e.g., laser-per-satellite

power systems, distributed power belt for orbit raising

• Spinoffs: powersats, power beaming, industrial lasers...


Where to go?

• Technology development -- only small niches– Most technology is being driven by other uses

– Some leverage in low-cost optics, SBC lasers

• Technology integration and demonstration– Integrated subscale module

• COTS fiber laser(s) or SBC laser array (~100 W)– Upgradeable to higher power as lasers become available

• Optics TBD: at least half-scale; full-scale if possible

• Full tracking system

– Full scale beam module is a bit much to bite off: ~$10-20 M• Higher power-per-module than originally conceived

• System integration and architecture studies– Many, many issues barely touched: siting, markets, safety...


Propellant(LH2) storage

Vehicle prep Payload handling

Launch catapultboosts vehicles to~100 m/s

Main Beam Module Array(100 - 10,000 units)

Power generation/energy storage

Array control

Secondary Beam Module array(s) for orbit raising, reentry,rendezvous propulsion, etc

Baseline:expandable

vehiclesdiscarded.

Vehicles couldalso be reused

Recovery area

Independent payloads go directly to LEO

Large spacecraft areassembled and serviced

at a LEO assembly facility(Crewed or robotic)

To GEO, Moon, Mars...

Supply vehiclesrendezvous with

Space Station andother future facilities

(Optionally) Vehiclesdiscard aeroshell,

drop tanks attop of atmosphere

Individual beams fromBeam Modulesadd incoherently atthe vehicle

Failed modules donot affect launch

Heat Exchanger(HX) vehiclewith side-facingheat exchanger

Laser Launch Architecture With Modular Ground-Based Laser Array

Modular Laser Launch Architecture: Analysis and Beam ... · A 100 MW launch system might have...

Documents

Transcript of Modular Laser Launch Architecture: Analysis and Beam ... · A 100 MW launch system might have...